Handover Minimized Service Region Partition for Mobile Edge Computing in Wireless Metropolitan Area Networks

Abstract

Along with the rapid proliferation of mobile applications, Mobile Edge Computing (MEC) attracts growing interests in recent years. As an essential component in MEC architecture, cloudlet handles the computation of applications offloaded from mobile devices, and pushes contents close to the mobile users, in order to improve the quality of experience of mobile users, as well as application deployment and delivery efficiency. Existing work mostly focuses on cloudlet placement and user-to-cloudlet association problem, assuming the capacities of cloudlets are given and fixed, with the goal of maximizing the tasks offloaded to network edge. We argue that the number of service handovers due to users’ movements between different MEC regions impacts heavily on QoE and service operation cost. Therefore, we propose a randomized algorithm to minimize the number of possible handovers between different MEC regions by carefully dividing a metropolitan area into disjoint clusters. The partition of MEC clusters is important since it is the base of other MEC resource allocation problems. Experiments on randomly generated traces and real traces exhibit that our algorithm could find sub-optimal partitions and significantly reduce the total number of handovers.